Genome Forensics Reveal Origins and the Mechanism that Drives t(12;21) Childhood B-ALL

At times, some of the most fundamental issues in human biology have appeared inaccessible to direct experimental investigation and have consequently been relegated to study using model systems. The origin of childhood precursor-B cell acute lymphocytic leukemia (B-ALL) was included in that group until the discovery 15 years ago of the prenatal origin of the t(12;21) translocation that generates the ETV6-RUNX1 (formerly TEL-AML1) fusion gene found in 25 percent of childhood ALL. In a succession of studies using blood samples from twin pairs, neonatal blood spots from archived Guthrie cards, and stored cord blood units, investigators showed that the ETV6-RUNX1 fusion gene developed in utero at an astounding frequency of one percent.¹  Clearly, 1 percent of children do not develop B-ALL. Thus, events acting in concert with t(12;21) are necessary for leukemogenesis.

During the Ham-Wasserman Lecture at the 2009 ASH annual meeting, Professor Mel Greaves, who had pioneered work in this area, made a compelling case for incorporating Darwinian evolutionary principles into a conceptual framework for understanding ALL leukemogenesis.¹  But at the time, the genetic events underlying clonal evolution of ETV6-RUNX1 expressing cells were speculative. Was the process merely stochastic, or was it connected mechanistically to t(12;21)? The expanding use of genome-scale technology over the past several years only added to the confusion by identifying many seemingly random genetic abnormalities. Now, the current study, led by Prof. Greaves from the Institute for Cancer Research in London and by Professor Peter Campbell from the University of Cambridge, Cambridge, UK, reports identification of the mechanism that underlies clonal expansion in B-ALL expressing ETV6-RUNX1.

In their detailed genetic analyses, Dr. Papaemmanuil and colleagues studied samples from 57 ETV6-RUNX1 positive ALL patients. Using both exome and low-coverage whole-genome sequencing, they found an average of 11 somatically acquired structural variations per case, mostly deletions, with half being highly recurrent. Notably, deletions mostly affected genes involved in B-cell differentiation. When the authors analyzed the nucleotide sequence of the structural variants, they observed a striking frequency of recombination signal sequence (RSS) motifs spanning breakpoint junctions. This finding suggested that recombination-activating endonucleases (RAG)-1 and -2 were involved in creation of the structural variants. RAG-1 and -2 are essential components of the process involved in generating antibody and T-cell receptor diversity through mediation of VDJ gene recombination. This process involves DNA cleavage and the addition of nontemplate nucleotide bases, which incidentally leaves a genomic fingerprint of RAG activity. The authors found that 40 percent of the analyzed structural variants had RAG recognition sequence (RSS) motifs and that 70 percent had inserted non-template sequence. Two other intriguing findings were the unexpectedly high incidence of structural variants that localized within promoter and enhancer regions, many of them coinciding with RSS motifs, and identification of mutations in two previously unrecognized tumor suppressor genes, ATF7IP and MGA. Interpreted as a whole, the mutational pattern suggests successive inactivation of genes that would otherwise promote differentiation, stalling the ETV-RUNX1 expressing cells in a stage of early B-lineage differentiation. Sustained high RAG activity would then drive clonal evolution, converting a pre-leukemic cell into a malignant clone.